shinys000114/linux
1
0
Fork 0
mirror of https://github.com/hardkernel/linux.git synced 2026-06-08 20:07:46 +09:00
Code Issues Packages Projects Releases Wiki Activity

mm: optimize for cma free time

Browse Source 
PD#165878: mm: optimize for cma free time

Fix bad parameter error and using high efficient free method.

Change-Id: I5fa1373644a373b3dd94775c5c29891f7f6f7420
Signed-off-by: tao zeng <tao.zeng@amlogic.com>
This commit is contained in:
tao zeng
2018-05-25 21:19:25 +08:00
committed by Yixun Lan
parent 241dc9bb02
commit 4b607551d1
1 changed files with 15 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

69
drivers/amlogic/memory_ext/aml_cma.c
View File

@@ -684,45 +684,16 @@ static int __aml_cma_free_check(struct page *page, int order, unsigned int *cnt)
return 0;
}
static inline unsigned long
__find_buddy_index(unsigned long page_idx, unsigned int order)
static int aml_cma_get_page_order(unsigned long pfn)
{
return page_idx ^ (1 << order);
}
int i, mask = 1;
static int aml_cma_get_page_order(struct page *page, int cur_order)
{
int order, ret = 0;
unsigned long page_idx, buddy_idx, combined_idx, flags;
struct page *buddy, *raw;
struct zone *zone;
/*
* same as __free_one_page, but we only need to find out freed buddy
* and it's order of it.
*/
buddy = page;
raw = page;
page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);
zone = page_zone(page);
spin_lock_irqsave(&zone->lock, flags);
for (order = cur_order; order < MAX_ORDER; order++) {
buddy_idx = __find_buddy_index(page_idx, order);
buddy = page + (buddy_idx - page_idx);
combined_idx = buddy_idx & page_idx;
page = page + (combined_idx - page_idx);
page_idx = combined_idx;
if (PageBuddy(buddy)) {
ret = page_private(buddy);
for (i = 0; i < (MAX_ORDER - 1); i++) {
if (pfn & (mask << i))
break;
}
}
spin_unlock_irqrestore(&zone->lock, flags);
WARN(ret >= MAX_ORDER, "%s page:%lx, buddy:%lx:%ld, corder:%d,ret:%d\n",
__func__, page_to_pfn(raw), page_to_pfn(buddy),
page_private(buddy), cur_order, ret);
return 0;
return i;
}
void aml_cma_free(unsigned long pfn, unsigned int nr_pages)
@@ -734,26 +705,7 @@ void aml_cma_free(unsigned long pfn, unsigned int nr_pages)
while (nr_pages) {
page = pfn_to_page(pfn);
batch = (1 << start_order);
if (__aml_cma_free_check(page, start_order, &count))
break;
__free_pages(page, start_order);
free_order = aml_cma_get_page_order(page, start_order);
pr_debug("pages:%4d, free:%2d, start:%2d, batch:%4d, pfn:%lx\n",
nr_pages, free_order,
start_order, batch, pfn);
nr_pages -= batch;
pfn += batch;
/*
* since pages are contigunous, and it's buddy already has large
* order, we can try to free same oder as free_order to get more
* quickly free speed.
*/
if (free_order < 0) {
start_order = 0;
continue;
}
free_order = aml_cma_get_page_order(pfn);
if (nr_pages >= (1 << free_order)) {
start_order = free_order;
} else {
@@ -763,6 +715,15 @@ void aml_cma_free(unsigned long pfn, unsigned int nr_pages)
start_order++;
start_order--;
}
batch = (1 << start_order);
if (__aml_cma_free_check(page, start_order, &count))
break;
__free_pages(page, start_order);
pr_debug("pages:%4d, free:%2d, start:%2d, batch:%4d, pfn:%lx\n",
nr_pages, free_order,
start_order, batch, pfn);
nr_pages -= batch;
pfn += batch;
}
WARN(count != 0, "%d pages are still in use!\n", count);
}